Static three-page website for the Philadelphia urban wind project.
index.html: project overview and figureswind-map.html: Leaflet map with OSM basemap and wind prediction XYZ tileshotspots.html: Leaflet map with Local Moran high-high hotspot polygons and popups
Run from /home/zimu:
conda run -n wind python wind_website/scripts/build_hotspots.pyWind tiles are generated from philly_predict/philly_wind_prediction.tif:
wind_website/scripts/build_wind_tiles.shThe tile script first colorizes the float raster, then warps it to EPSG:3857,
then writes assets/tiles/wind/{z}/{x}/{y}.png. The explicit source CRS string
keeps the projection step offline when PROJ datum-grid downloads are unavailable.
The current wind color ramp is a viridis-style scale from purple/blue at low
wind speed to yellow at high wind speed, stretched across roughly 0-2 m/s.
Hotspot popups link to Google Street View using representative point coordinates.
assets/data/hotspot_points.geojson stores those clickable points.
The first point for each hotspot region is the polygon centroid when that centroid falls inside the hotspot. If the centroid falls outside a concave region, the point is moved to the nearest hotspot analysis-cell center. Regions larger than 40 analysis cells get multiple distributed Street View points, up to 12 per region.
cd /home/zimu/wind_website
python -m http.server 8050Open http://localhost:8050.
The site vendors Leaflet in assets/vendor/leaflet, so the map library does
not depend on a CDN. The hotspot page also loads assets/data/hotspots-data.js
and assets/data/hotspot-points-data.js as direct-script fallbacks, which makes
the hotspot polygons and Street View points available even if the page is opened
from file:// and the browser blocks fetch() for local GeoJSON files.
Create a GitHub repository, push this folder, and enable Pages from the repository settings. Use the repository root as the Pages source.